Between 15 - 20% of children and adolescents are estimated to be affected by anxiety disorders, making anxiety the most frequently diagnosed form of psychopathology in this population. Therefore, there is a critical need to identify markers of risk that predict normative versus anxious trajectories of development that could inform etiopathogenesis, and predict subsequent outcomes; such markers could also be critical for identifying individuals for intervention and prevention efforts. Under the guidance of Dr. Hajcak Proudfit, the current proposal aims to investigate error related brain activity, a promising biomarker of anxiety, in a large sample of children and adolescent females during a core risk period. ERPs and fMRI data, as well as diagnostic interviews, and self-report measures are currently being collected in a large (N=150) and longitudinal ongoing NIMH-funded R01 study on girls ranging from 8 to 14 years of age, oversampled based on risk for anxiety; anxiety symptoms and diagnoses will be assessed approximately 2 years after the initial visit. Using multiple measures of error processing (2 ERP tasks and 1 fMRI) and anxiety (self-report and interview), I will model error sensitivity and trait anxiety as latent variables and examine both cross-sectional and prospective relationships, allowing for analyses to determine whether different neural measures capture both unique and overlapping variance related to anxiety and whether error sensitivity can predict increased anxiety across adolescent development, controlling for initial anxiety symptoms. The current training plan will expand my methodological skills to fMRI, so that I can examine multiple neural measures of error-processing; additionally, I will learn advanced statistical approaches to model cross-sectional and prospective relationships between error processing and anxiety as latent traits. The current proposal builds on prior work to address a gap in the literature regarding the specificity of different neural measures of error processing to anxiety, as well as the validity of error processing as a risk marker that can predict increased anxiety prospectively.